RainFall Prediction¶

Problem Statement¶

Build an efficient Classification Model that should predict whether it Rains Tomorrow or not, using the dataset¶

Features description¶

Date---The date of observation¶

Location---The common name of the location of the weather station¶

MinTemp---The minimum temperature in degrees celsius¶

MaxTemp---The maximum temperature in degrees celsius¶

Rainfall---The amount of rainfall recorded for the day in mm¶

Evaporation---The so-called Class A pan evaporation (mm) in the 24 hours to 9am¶

Sunshine---The number of hours of bright sunshine in the day.¶

WindGustDir---The direction of the strongest wind gust in the 24 hours to midnight¶

WindGustSpeed---The speed (km/h) of the strongest wind gust in the 24 hours to midnight¶

WindDir9am---Direction of the wind at 9am¶

WindDir3pm---Direction of the wind at 3pm¶

WindSpeed9am---Wind speed (km/hr) averaged over 10 minutes prior to 9am¶

WindSpeed3pm---Wind speed (km/hr) averaged over 10 minutes prior to 3pm¶

Humidity9am---Humidity (percent) at 9am¶

Humidity3pm---Humidity (percent) at 3pm¶

Pressure9am---Atmospheric pressure (hpa) reduced to mean sea level at 9am¶

Pressure3pm---Atmospheric pressure (hpa) reduced to mean sea level at 3pm¶

Cloud9am---Fraction of sky obscured by cloud at 9am. This is measured in "oktas", which are a unit of eigths. It records how many eigths of the sky are obscured by cloud. A 0 measure indicates completely clear sky whilst an 8 indicates that it is completely overcast.¶

Cloud3pm---Fraction of sky obscured by cloud (in "oktas": eighths) at 3pm. See Cload9am for a description of the values¶

Temp9am---Temperature (degrees C) at 9am¶

RISK_MM---The amount of next day rain in mm. Used to create response variable RainTomorrow. A kind of measure of the "risk".¶

RainTomorrow---The target variable. Did it rain tomorrow?¶

Temp3pm---Temperature (degrees C) at 3pm¶

RainToday---Boolean: 1 if precipitation (mm) in the 24 hours to 9am exceeds 1mm, otherwise 0¶

In [76]:
# Importing required libraries

import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
In [77]:
# Importing the data

df = pd.read_csv(r"C:\Users\shinu\Downloads\Rainfall Prediction\weatherAUS.csv")
In [78]:
df
Out[78]:
Date Location MinTemp MaxTemp Rainfall Evaporation Sunshine WindGustDir WindGustSpeed WindDir9am ... Humidity3pm Pressure9am Pressure3pm Cloud9am Cloud3pm Temp9am Temp3pm RainToday RISK_MM RainTomorrow
0 2008-12-01 Albury 13.4 22.9 0.6 NaN NaN W 44.0 W ... 22.0 1007.7 1007.1 8.0 NaN 16.9 21.8 No 0.0 No
1 2008-12-02 Albury 7.4 25.1 0.0 NaN NaN WNW 44.0 NNW ... 25.0 1010.6 1007.8 NaN NaN 17.2 24.3 No 0.0 No
2 2008-12-03 Albury 12.9 25.7 0.0 NaN NaN WSW 46.0 W ... 30.0 1007.6 1008.7 NaN 2.0 21.0 23.2 No 0.0 No
3 2008-12-04 Albury 9.2 28.0 0.0 NaN NaN NE 24.0 SE ... 16.0 1017.6 1012.8 NaN NaN 18.1 26.5 No 1.0 No
4 2008-12-05 Albury 17.5 32.3 1.0 NaN NaN W 41.0 ENE ... 33.0 1010.8 1006.0 7.0 8.0 17.8 29.7 No 0.2 No
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
142188 2017-06-20 Uluru 3.5 21.8 0.0 NaN NaN E 31.0 ESE ... 27.0 1024.7 1021.2 NaN NaN 9.4 20.9 No 0.0 No
142189 2017-06-21 Uluru 2.8 23.4 0.0 NaN NaN E 31.0 SE ... 24.0 1024.6 1020.3 NaN NaN 10.1 22.4 No 0.0 No
142190 2017-06-22 Uluru 3.6 25.3 0.0 NaN NaN NNW 22.0 SE ... 21.0 1023.5 1019.1 NaN NaN 10.9 24.5 No 0.0 No
142191 2017-06-23 Uluru 5.4 26.9 0.0 NaN NaN N 37.0 SE ... 24.0 1021.0 1016.8 NaN NaN 12.5 26.1 No 0.0 No
142192 2017-06-24 Uluru 7.8 27.0 0.0 NaN NaN SE 28.0 SSE ... 24.0 1019.4 1016.5 3.0 2.0 15.1 26.0 No 0.0 No

142193 rows × 24 columns

In [79]:
df.head()
Out[79]:
Date Location MinTemp MaxTemp Rainfall Evaporation Sunshine WindGustDir WindGustSpeed WindDir9am ... Humidity3pm Pressure9am Pressure3pm Cloud9am Cloud3pm Temp9am Temp3pm RainToday RISK_MM RainTomorrow
0 2008-12-01 Albury 13.4 22.9 0.6 NaN NaN W 44.0 W ... 22.0 1007.7 1007.1 8.0 NaN 16.9 21.8 No 0.0 No
1 2008-12-02 Albury 7.4 25.1 0.0 NaN NaN WNW 44.0 NNW ... 25.0 1010.6 1007.8 NaN NaN 17.2 24.3 No 0.0 No
2 2008-12-03 Albury 12.9 25.7 0.0 NaN NaN WSW 46.0 W ... 30.0 1007.6 1008.7 NaN 2.0 21.0 23.2 No 0.0 No
3 2008-12-04 Albury 9.2 28.0 0.0 NaN NaN NE 24.0 SE ... 16.0 1017.6 1012.8 NaN NaN 18.1 26.5 No 1.0 No
4 2008-12-05 Albury 17.5 32.3 1.0 NaN NaN W 41.0 ENE ... 33.0 1010.8 1006.0 7.0 8.0 17.8 29.7 No 0.2 No

5 rows × 24 columns

In [81]:
df.tail()
Out[81]:
Date Location MinTemp MaxTemp Rainfall Evaporation Sunshine WindGustDir WindGustSpeed WindDir9am ... Humidity3pm Pressure9am Pressure3pm Cloud9am Cloud3pm Temp9am Temp3pm RainToday RISK_MM RainTomorrow
142188 2017-06-20 Uluru 3.5 21.8 0.0 NaN NaN E 31.0 ESE ... 27.0 1024.7 1021.2 NaN NaN 9.4 20.9 No 0.0 No
142189 2017-06-21 Uluru 2.8 23.4 0.0 NaN NaN E 31.0 SE ... 24.0 1024.6 1020.3 NaN NaN 10.1 22.4 No 0.0 No
142190 2017-06-22 Uluru 3.6 25.3 0.0 NaN NaN NNW 22.0 SE ... 21.0 1023.5 1019.1 NaN NaN 10.9 24.5 No 0.0 No
142191 2017-06-23 Uluru 5.4 26.9 0.0 NaN NaN N 37.0 SE ... 24.0 1021.0 1016.8 NaN NaN 12.5 26.1 No 0.0 No
142192 2017-06-24 Uluru 7.8 27.0 0.0 NaN NaN SE 28.0 SSE ... 24.0 1019.4 1016.5 3.0 2.0 15.1 26.0 No 0.0 No

5 rows × 24 columns

EDA¶

In [170]:
df['WindGustDir'].value_counts().plot(kind="pie",autopct="%.2f%%",figsize=(10,7))
plt.show()
In [168]:
df['WindDir9am'].value_counts().plot(kind="pie",autopct="%.2f%%",figsize=(10,7))
plt.show()
In [166]:
df['WindDir3pm'].value_counts().plot(kind="pie",autopct="%.2f%%",figsize=(10,7))
plt.show()
In [174]:
df['Humidity3pm'].value_counts().plot(kind="bar",figsize=(20,15))
plt.show()
In [175]:
df['RainTomorrow'].value_counts().plot(kind="pie")
plt.show()
In [89]:
ax=pd.crosstab(df['WindGustDir'],df['WindDir9am']).plot(kind="bar",stacked=True,figsize = (25,15))
for i in ax.containers:
    ax.bar_label(i)
In [90]:
ax=pd.crosstab(df['WindDir9am'],df['WindDir3pm']).plot(kind="bar",stacked=True,figsize = (25,15))
for i in ax.containers:
    ax.bar_label(i)
In [103]:
ax=pd.crosstab(df['RainToday'],df['RainTomorrow']).plot(kind="bar",stacked=True)
for i in ax.containers:
    ax.bar_label(i)
In [93]:
# Pairplot for whole data tells you the distribution for single column and that single column with all other different columns

sns.pairplot(df,diag_kind='kde')
Out[93]:
<seaborn.axisgrid.PairGrid at 0x1e50240b7d0>
In [94]:
# Finding the distribution for each single columnb

df.hist(figsize=(20,15))
plt.show()
In [35]:
# Checking Correlation between two variabless using Heatmap

k=df.select_dtypes(include=['int','float'])
cor=k.corr()
plt.figure(figsize=(20,15))
sns.heatmap(cor,annot=True)
plt.show()
In [104]:
# sorting the data based on date (Time based splitting)

df = df.sort_values(by='Date')
In [105]:
#After sorting the data index's got changed to get that in order we doing reset index

df.reset_index(inplace=True)
df.drop("index",axis = 1,inplace=True)
In [106]:
#Removing unwanted features, RISK_MM is same as target label hence removing with date and loaction

df.drop(['Date', 'Location','RISK_MM'], axis=1,inplace = True)
In [108]:
df.head()
Out[108]:
MinTemp MaxTemp Rainfall Evaporation Sunshine WindGustDir WindGustSpeed WindDir9am WindDir3pm WindSpeed9am ... Humidity9am Humidity3pm Pressure9am Pressure3pm Cloud9am Cloud3pm Temp9am Temp3pm RainToday RainTomorrow
0 8.0 24.3 0.0 3.4 6.3 NW 30.0 SW NW 6.0 ... 68.0 29.0 1019.7 1015.0 7.0 7.0 14.4 23.6 No Yes
1 14.0 26.9 3.6 4.4 9.7 ENE 39.0 E W 4.0 ... 80.0 36.0 1012.4 1008.4 5.0 3.0 17.5 25.7 Yes Yes
2 13.7 23.4 3.6 5.8 3.3 NW 85.0 N NNE 6.0 ... 82.0 69.0 1009.5 1007.2 8.0 7.0 15.4 20.2 Yes Yes
3 13.3 15.5 39.8 7.2 9.1 NW 54.0 WNW W 30.0 ... 62.0 56.0 1005.5 1007.0 2.0 7.0 13.5 14.1 Yes Yes
4 7.6 16.1 2.8 5.6 10.6 SSE 50.0 SSE ESE 20.0 ... 68.0 49.0 1018.3 1018.5 7.0 7.0 11.1 15.4 Yes No

5 rows × 21 columns

In [110]:
df.dtypes
Out[110]:
MinTemp          float64
MaxTemp          float64
Rainfall         float64
Evaporation      float64
Sunshine         float64
WindGustDir       object
WindGustSpeed    float64
WindDir9am        object
WindDir3pm        object
WindSpeed9am     float64
WindSpeed3pm     float64
Humidity9am      float64
Humidity3pm      float64
Pressure9am      float64
Pressure3pm      float64
Cloud9am         float64
Cloud3pm         float64
Temp9am          float64
Temp3pm          float64
RainToday         object
RainTomorrow      object
dtype: object

Chi Square Test¶

In [111]:
# Both variables categorical - chi square test of independence

pd.crosstab(df['WindGustDir'],df['WindDir9am'])
Out[111]:
WindDir9am E ENE ESE N NE NNE NNW NW S SE SSE SSW SW W WNW WSW
WindGustDir
E 2251 1169 1353 168 424 219 136 196 312 801 517 257 308 134 218 164
ENE 1225 1560 699 256 808 383 166 195 281 535 413 230 248 201 172 182
ESE 1000 429 1478 157 233 170 104 133 388 1185 707 267 240 146 130 137
N 310 431 177 2966 785 1441 677 420 95 180 125 102 126 198 223 125
NE 573 918 344 472 1255 834 285 322 180 368 228 152 173 153 182 128
NNE 321 435 204 961 900 1433 354 340 105 200 138 103 100 154 143 113
NNW 229 257 136 1422 381 570 1329 719 99 135 114 69 87 151 256 90
NW 245 262 139 1196 343 520 1385 1484 120 151 132 107 135 349 710 179
S 243 197 326 232 160 185 178 239 1894 593 1120 1268 850 308 255 475
SE 663 350 1042 170 252 153 178 161 927 2014 1761 419 311 162 162 161
SSE 361 204 548 189 142 128 120 136 1409 1338 2093 721 465 262 215 279
SSW 304 260 247 282 228 233 204 224 1112 318 542 1629 1247 404 306 629
SW 296 291 232 383 253 267 280 289 609 277 348 973 1683 695 342 1106
W 242 260 113 756 297 378 725 910 163 148 162 237 527 2051 1349 966
WNW 174 207 110 797 274 450 925 1289 121 123 134 128 210 828 1477 340
WSW 224 276 162 528 305 339 424 469 334 164 220 487 1019 1305 705 1456
In [112]:
# Null-there is no association between both variables
# Alt-There is association between both variables
In [113]:
from scipy.stats import chi2_contingency
chi2_contingency(pd.crosstab(df['WindGustDir'],df['WindDir9am']))

#Since Pvalue = 0.0 Not Reject Null
Out[113]:
Chi2ContingencyResult(statistic=110952.90285769245, pvalue=0.0, dof=225, expected_freq=array([[596.13562527, 516.63710926, 503.14645199, 752.65478147,
        484.56238332, 530.1965964 , 514.15923343, 518.01370694,
        560.89472466, 587.11891047, 602.53680448, 492.06484067,
        531.98617339, 516.29295984, 471.14055594, 449.45914248],
       [521.99009079, 452.37935822, 440.56662784, 659.04187078,
        424.29398905, 464.25235762, 450.20967304, 453.58473887,
        491.13234614, 514.09484753, 527.59511082, 430.8633136 ,
        465.81935247, 452.07801305, 412.54152771, 393.55678246],
       [477.07434298, 413.45341397, 402.65713511, 602.333211  ,
        387.78471014, 424.30477589, 411.47042397, 414.55507508,
        448.8718186 , 469.85846272, 482.19706713, 393.78876318,
        425.73693533, 413.17799869, 377.04351434, 359.69235188],
       [579.13674225, 501.90513651, 488.79916705, 731.19273484,
        470.74502545, 515.07797316, 499.49791763, 503.24248033,
        544.9007404 , 570.37714021, 585.35539102, 478.03354928,
        516.81652013, 501.57080056, 457.70592318, 436.64275798],
       [453.78725526, 393.27180903, 383.00252118, 572.931952  ,
        368.85605323, 403.5934912 , 391.3856133 , 394.31969554,
        426.96136048, 446.92359859, 458.65992756, 374.56703474,
        404.95574367, 393.0098374 , 358.63915971, 342.1349471 ],
       [414.88330754, 359.5559527 , 350.16706825, 523.81352822,
        337.23340088, 368.99273963, 357.83146372, 360.51400214,
        390.35724202, 408.60808374, 419.33823741, 342.4547703 ,
        370.2382039 , 359.31644035, 327.8924189 , 312.80314031],
       [417.64735356, 361.95139543, 352.49996011, 527.3032919 ,
        339.48012574, 371.45105235, 360.21541751, 362.9158276 ,
        392.95788987, 411.33032281, 422.13196317, 344.7362811 ,
        372.70481418, 361.71028738, 330.07691203, 314.88710527],
       [515.2872792 , 446.57040961, 434.90936508, 650.57919386,
        418.84568128, 458.29094927, 444.42858511, 447.76031212,
        484.8257751 , 507.4934178 , 520.82032584, 425.33064992,
        459.83782253, 446.27293399, 407.24413187, 388.50316744],
       [588.94910562, 510.40895818, 497.08093316, 743.58139591,
        478.72089869, 523.80498333, 507.96095358, 511.76896073,
        554.13304026, 580.0410889 , 595.27311749, 486.1329126 ,
        525.57298664, 510.06895754, 465.4608738 , 444.04083359],
       [614.03282325, 532.14760089, 518.25192679, 775.25100129,
        499.10992676, 546.11417128, 529.59533422, 533.56552682,
        577.73391948, 604.74540842, 620.62617881, 506.83762307,
        547.95747499, 531.79311941, 485.28514896, 462.95281559],
       [594.96090571, 515.6190461 , 502.15497295, 751.17163191,
        483.60752525, 529.1518135 , 513.14605307, 516.99293111,
        559.78944933, 585.96195886, 601.34947103, 491.09519858,
        530.93786402, 515.27557485, 470.21214636, 448.57345737],
       [564.48729835, 489.20929008, 476.43484019, 712.69698735,
        458.83738371, 502.04891573, 486.86296255, 490.51280537,
        531.1173068 , 555.94927317, 570.54864447, 465.94154207,
        503.74348561, 488.88341126, 446.12810959, 425.5977437 ],
       [575.19797667, 498.49163063, 485.47479615, 726.2198216 ,
        467.54344253, 511.57487753, 496.10078348, 499.81987905,
        541.19481721, 566.49794954, 581.37433181, 474.7823964 ,
        513.30160047, 498.15956853, 454.59302047, 433.67310792],
       [641.53508114, 555.98225598, 541.4642008 , 809.9741499 ,
        521.46483908, 570.57438287, 553.31567442, 557.46369018,
        603.61036557, 631.83168712, 648.42375018, 529.53865548,
        572.50024733, 555.61189743, 507.02085561, 483.68826693],
       [524.27042876, 454.35559846, 442.49126362, 661.92092582,
        426.14753706, 466.28046562, 452.17643492, 455.56624487,
        493.27788061, 516.34069476, 529.89993458, 432.74556001,
        467.85430596, 454.05293686, 414.34373454, 395.27605355],
       [581.62438367, 504.06103496, 490.89876973, 734.33352216,
        472.76707782, 517.29045461, 501.64347604, 505.40412325,
        547.24132346, 572.82715537, 587.86974421, 480.086909  ,
        519.03646939, 503.72526289, 459.671967  , 438.51832645]]))
In [115]:
pd.crosstab(df['WindDir9am'],df['WindDir3pm'])
Out[115]:
WindDir3pm E ENE ESE N NE NNE NNW NW S SE SSE SSW SW W WNW WSW
WindDir9am
E 1520 1049 1110 412 655 416 362 349 342 752 467 314 348 262 244 308
ENE 842 1110 553 592 818 564 415 325 237 377 274 265 338 338 244 368
ESE 1091 775 1265 232 417 219 207 178 365 1066 610 249 297 172 147 207
N 247 292 188 2224 698 905 1687 1363 250 219 199 255 408 839 955 524
NE 379 628 264 890 1045 897 565 467 187 277 170 194 284 364 380 355
NNE 233 353 190 1522 831 1097 791 624 174 148 136 175 242 506 470 387
NNW 182 187 211 597 396 379 1150 1259 181 273 121 177 300 827 1068 450
NW 244 226 222 410 463 377 671 1356 199 292 166 207 355 1070 1347 596
S 370 314 515 104 249 119 133 141 1753 969 1458 922 651 187 161 345
SE 769 667 1169 218 458 243 242 185 694 1934 1159 376 306 166 182 189
SSE 559 469 882 171 303 149 179 176 1097 1695 1674 535 447 170 145 245
SSW 286 278 381 95 211 100 84 120 1357 523 780 1327 982 204 141 449
SW 310 292 327 125 263 142 103 158 1071 522 617 1195 1387 471 194 885
W 388 245 289 177 248 129 168 327 438 256 328 491 792 1696 817 1358
WNW 327 214 237 186 259 132 229 597 266 244 224 260 425 1298 1366 868
WSW 208 213 228 112 167 106 101 164 563 285 348 655 1127 835 323 1331
In [116]:
chi2_contingency(pd.crosstab(df['WindDir9am'],df['WindDir3pm']))

# Since pvalue=0.0 , Not Reject Null
Out[116]:
Chi2ContingencyResult(statistic=79917.66217942107, pvalue=0.0, dof=225, expected_freq=array([[544.49885921, 500.48719781, 549.70085963, 552.16496508,
        512.05480399, 408.90461156, 485.08653868, 533.13659515,
        627.93620797, 672.97457998, 597.61402134, 519.99469936,
        594.73923164, 643.74755134, 560.17330783, 606.78596944],
       [468.11013036, 430.27294447, 472.58233274, 474.70074439,
        440.21771028, 351.53864473, 417.03287164, 458.34189886,
        539.8419027 , 578.56176012, 513.77367042, 447.04370338,
        511.30219016, 553.43504413, 481.58558226, 521.65886935],
       [458.14904012, 421.11700583, 462.52607684, 464.59941002,
        430.85015326, 344.05812265, 408.15867346, 448.58867046,
        528.35440529, 566.25032841, 502.84088866, 437.5308935 ,
        500.42199995, 541.65829319, 471.33774285, 510.5582955 ],
       [687.68189256, 632.09679427, 694.25182642, 697.36390035,
        646.70625245, 516.43137978, 612.64633219, 673.33177387,
        793.06017377, 849.94196953, 754.7643751 , 656.73404623,
        751.13362218, 813.02931483, 707.47814063, 766.34820585],
       [448.92128168, 412.63512403, 453.21015879, 455.24173216,
        422.17223234, 337.12831386, 399.93779048, 439.55347115,
        517.71261321, 554.84525977, 492.71297427, 428.71841319,
        490.34280534, 530.74854232, 461.8443456 , 500.27494181],
       [481.49343566, 442.57448165, 486.09349865, 488.27247586,
        452.8035691 , 361.58916212, 428.95587411, 471.44592965,
        555.27602498, 595.10288616, 528.4624999 , 459.82471788,
        525.92035983, 569.2577954 , 495.35415178, 536.57313729],
       [474.09900671, 435.77774193, 478.62842525, 480.7739393 ,
        445.84973843, 356.03613653, 422.36827914, 464.20580305,
        546.74849623, 585.9637252 , 520.34675394, 452.76306146,
        517.84365421, 560.5155447 , 487.74685995, 528.332834  ],
       [501.17117221, 460.66167331, 505.95923118, 508.22725911,
        471.30880444, 376.36663517, 446.48649874, 490.71304341,
        577.96911802, 619.42362856, 550.0597743 , 478.61689444,
        547.41374171, 592.52229725, 515.59834989, 558.50187827],
       [512.78225899, 471.33423982, 517.68124726, 520.00182065,
        482.22804268, 385.08626213, 456.83065613, 502.08183725,
        591.35945242, 633.77437717, 562.80350764, 489.70544583,
        560.09617202, 606.24979835, 527.54368417, 571.44119748],
       [547.37107542, 503.12725373, 552.60051624, 555.07761978,
        514.75587872, 411.06157191, 487.64535656, 535.94887573,
        631.24855385, 676.52450201, 600.76641854, 522.73765681,
        597.8764644 , 647.14330161, 563.12820631, 609.9867484 ],
       [543.64330545, 499.70079817, 548.83713212, 551.29736581,
        511.25022854, 408.26211273, 484.32433761, 532.29889455,
        626.94955175, 671.9171564 , 596.67500941, 519.17764821,
        593.80473677, 642.73605125, 559.2931253 , 605.83254592],
       [447.21017415, 411.06232475, 451.48270379, 453.50653361,
        420.56308144, 335.8433162 , 398.41338834, 437.87806995,
        515.73930078, 552.7304126 , 490.83495041, 427.08431088,
        488.47381561, 528.72554216, 460.08398055, 498.36809477],
       [492.67674556, 452.85384834, 497.38365099, 499.61323777,
        463.32051962, 369.98753966, 438.91893096, 482.39587318,
        568.17303127, 608.92492299, 540.73672728, 470.50474369,
        538.1355427 , 582.47954645, 506.85939481, 549.03574474],
       [497.87117912, 457.62841757, 502.62771082, 504.88080478,
        468.20544199, 373.8884254 , 443.54658032, 487.48191253,
        574.16344403, 615.34499474, 546.43787114, 475.46541141,
        543.80926152, 588.62079694, 512.20336014, 554.82438755],
       [435.8435313 , 400.61444386, 440.00746699, 441.97985757,
        409.87372189, 327.30726034, 388.28700268, 426.74861915,
        502.63086815, 538.68178501, 478.3595062 , 416.22920268,
        476.05838384, 515.28704109, 448.39012698, 485.70118227],
       [413.47691149, 380.05571048, 417.42716231, 419.29833376,
        388.83982085, 310.51050525, 368.36088897, 404.84873207,
        476.83685557, 511.03771135, 453.81105145, 394.86915105,
        451.6280181 , 488.84353898, 425.37964094, 460.77596737]]))

Data Preprocessing¶

In [117]:
#Checking for null valuess

df.isna().sum()
Out[117]:
MinTemp            637
MaxTemp            322
Rainfall          1406
Evaporation      60843
Sunshine         67816
WindGustDir       9330
WindGustSpeed     9270
WindDir9am       10013
WindDir3pm        3778
WindSpeed9am      1348
WindSpeed3pm      2630
Humidity9am       1774
Humidity3pm       3610
Pressure9am      14014
Pressure3pm      13981
Cloud9am         53657
Cloud3pm         57094
Temp9am            904
Temp3pm           2726
RainToday         1406
RainTomorrow         0
dtype: int64
In [120]:
df.columns
Out[120]:
Index(['MinTemp', 'MaxTemp', 'Rainfall', 'Evaporation', 'Sunshine',
       'WindGustDir', 'WindGustSpeed', 'WindDir9am', 'WindDir3pm',
       'WindSpeed9am', 'WindSpeed3pm', 'Humidity9am', 'Humidity3pm',
       'Pressure9am', 'Pressure3pm', 'Cloud9am', 'Cloud3pm', 'Temp9am',
       'Temp3pm', 'RainToday', 'RainTomorrow'],
      dtype='object')
In [121]:
#Creating a list to fill the Median for numerical columns

Numerical_cols = ['MinTemp', 'MaxTemp', 'Rainfall', 'Evaporation',
       'Sunshine', 'WindGustSpeed','WindSpeed9am', 'WindSpeed3pm', 'Humidity9am', 'Humidity3pm',
       'Pressure9am', 'Pressure3pm', 'Cloud9am', 'Cloud3pm', 'Temp9am',
       'Temp3pm']
In [122]:
#Filling the values and making permanent change 

for c in Numerical_cols:
    p = df[c].median()
    df[c] = df[c].fillna(p)
In [123]:
##Creating a list to fill the Mean for categorical columns


Categorical_cols = ["WindGustDir","WindDir9am","WindDir3pm","RainToday"]
In [124]:
#Filling the values and making permanent change 


for x in Categorical_cols:
    q = df[x].mode()[0]
    df[x] = df[x].fillna(q)
In [125]:
df.dtypes
Out[125]:
MinTemp          float64
MaxTemp          float64
Rainfall         float64
Evaporation      float64
Sunshine         float64
WindGustDir       object
WindGustSpeed    float64
WindDir9am        object
WindDir3pm        object
WindSpeed9am     float64
WindSpeed3pm     float64
Humidity9am      float64
Humidity3pm      float64
Pressure9am      float64
Pressure3pm      float64
Cloud9am         float64
Cloud3pm         float64
Temp9am          float64
Temp3pm          float64
RainToday         object
RainTomorrow      object
dtype: object

Encoding¶

In [126]:
import warnings
warnings.filterwarnings("ignore")
In [127]:
#After EDA we having some categorical columns i.e., WindGustDir,WindDir9am,WindDir3pm,RainToday,RainTomorrow 


#For Rain Tomorrow column we use LabelEncoder
from sklearn.preprocessing import LabelEncoder
le = LabelEncoder()
df["RainTomorrow"] = le.fit_transform(df[["RainTomorrow"]])


from sklearn.preprocessing import LabelEncoder
le = LabelEncoder()
df["WindGustDir"] = le.fit_transform(df[["WindGustDir"]])


from sklearn.preprocessing import LabelEncoder
le = LabelEncoder()
df["WindDir9am"] = le.fit_transform(df[["WindDir9am"]])



from sklearn.preprocessing import LabelEncoder
le = LabelEncoder()
df["WindDir3pm"] = le.fit_transform(df[["WindDir3pm"]])


from sklearn.preprocessing import LabelEncoder
le = LabelEncoder()
df["RainToday"] = le.fit_transform(df[["RainToday"]])
In [128]:
df
Out[128]:
MinTemp MaxTemp Rainfall Evaporation Sunshine WindGustDir WindGustSpeed WindDir9am WindDir3pm WindSpeed9am ... Humidity9am Humidity3pm Pressure9am Pressure3pm Cloud9am Cloud3pm Temp9am Temp3pm RainToday RainTomorrow
0 8.0 24.3 0.0 3.4 6.3 7 30.0 12 7 6.0 ... 68.0 29.0 1019.7 1015.0 7.0 7.0 14.4 23.6 0 1
1 14.0 26.9 3.6 4.4 9.7 1 39.0 0 13 4.0 ... 80.0 36.0 1012.4 1008.4 5.0 3.0 17.5 25.7 1 1
2 13.7 23.4 3.6 5.8 3.3 7 85.0 3 5 6.0 ... 82.0 69.0 1009.5 1007.2 8.0 7.0 15.4 20.2 1 1
3 13.3 15.5 39.8 7.2 9.1 7 54.0 14 13 30.0 ... 62.0 56.0 1005.5 1007.0 2.0 7.0 13.5 14.1 1 1
4 7.6 16.1 2.8 5.6 10.6 10 50.0 10 2 20.0 ... 68.0 49.0 1018.3 1018.5 7.0 7.0 11.1 15.4 1 0
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
142188 -0.1 12.7 0.0 4.8 8.5 11 19.0 11 10 2.0 ... 92.0 54.0 1020.8 1018.5 8.0 8.0 4.0 10.7 0 0
142189 7.6 19.3 0.0 3.4 9.4 13 35.0 13 13 13.0 ... 73.0 32.0 1018.6 1015.4 1.0 1.0 9.4 18.8 0 0
142190 2.7 20.0 0.2 4.8 8.5 13 37.0 3 14 0.0 ... 88.0 35.0 1018.8 1015.2 5.0 5.0 9.6 19.4 0 0
142191 3.1 20.0 0.0 4.8 8.5 12 28.0 9 8 2.0 ... 95.0 45.0 1017.6 1015.2 5.0 5.0 7.8 17.8 0 0
142192 6.8 19.5 0.0 3.4 9.4 15 46.0 7 15 19.0 ... 67.0 42.0 1018.4 1015.2 1.0 1.0 11.9 17.9 0 0

142193 rows × 21 columns

Splitting the data¶

In [129]:
X = df.drop("RainTomorrow",axis = 1)
Y = df["RainTomorrow"]
In [130]:
from sklearn.model_selection import train_test_split
X_train,X_test,Y_train,Y_test = train_test_split(X,Y,test_size=0.30)

Model Building¶

In [131]:
from sklearn.linear_model import LogisticRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import GradientBoostingClassifier
from xgboost import XGBClassifier

Logistic Regression¶

In [132]:
from sklearn.linear_model import LogisticRegression
lr = LogisticRegression()
lr.fit(X,Y)
lr.score(X,Y)
Out[132]:
0.838318341971827
In [133]:
from sklearn.model_selection import cross_val_score
cv=cross_val_score(lr,X,Y)
print(cv)
print()
print(cv.mean())

[0.84310278 0.83835578 0.84292697 0.8320557  0.83497433]

0.8382831114316976

Model Evaluation¶

In [135]:
RocCurveDisplay.from_predictions(Y_test,lr_pred)
plt.show()
In [136]:
from sklearn.metrics import confusion_matrix,classification_report,f1_score,RocCurveDisplay
lr_pred = lr.predict(X_test)
In [137]:
print(confusion_matrix(Y_test,lr_pred))

[[31256  1859]
 [ 5058  4485]]
In [138]:
print(classification_report(Y_test,lr_pred))

              precision    recall  f1-score   support

           0       0.86      0.94      0.90     33115
           1       0.71      0.47      0.56      9543

    accuracy                           0.84     42658
   macro avg       0.78      0.71      0.73     42658
weighted avg       0.83      0.84      0.83     42658
In [ ]:

DecisionTree Classifier¶

In [141]:
from sklearn.tree import DecisionTreeClassifier
dt = DecisionTreeClassifier(max_depth=10)
dt.fit(X,Y)
dt.score(X,Y)
Out[141]:
0.8600423368238943
In [142]:
cv1=cross_val_score(dt,X,Y)
print(cv1)
print()
print(cv1.mean())

[0.8371954  0.83431204 0.83902388 0.83328645 0.83170406]

0.8351043636590996

Model Evaluation¶

In [143]:
RocCurveDisplay.from_predictions(Y_test,dt_pred)
plt.show()
In [144]:
from sklearn.metrics import confusion_matrix,classification_report
dt_pred = dt.predict(X_test)
In [145]:
print(confusion_matrix(Y_test,dt_pred))

[[31620  1495]
 [ 4460  5083]]
In [146]:
print(classification_report(Y_test,dt_pred))

              precision    recall  f1-score   support

           0       0.88      0.95      0.91     33115
           1       0.77      0.53      0.63      9543

    accuracy                           0.86     42658
   macro avg       0.82      0.74      0.77     42658
weighted avg       0.85      0.86      0.85     42658
In [ ]:

Random forest Classifier¶

In [147]:
from sklearn.ensemble import RandomForestClassifier
rfc = RandomForestClassifier(max_depth=10)
rfc.fit(X,Y)
rfc.score(X,Y)
Out[147]:
0.8631718860984718
In [148]:
cv2=cross_val_score(rfc,X,Y)
print(cv2)
print()
print(cv2.mean())

[0.84988924 0.8483069  0.8508738  0.84028413 0.84123356]

0.8461175253538282

Model Evaluation¶

In [149]:
from sklearn.metrics import confusion_matrix,classification_report
rfc_pred = rfc.predict(X_test)
In [150]:
RocCurveDisplay.from_predictions(Y_test,rfc_pred)
plt.show()
In [151]:
print(confusion_matrix(Y_test,rfc_pred))

[[32136   979]
 [ 4850  4693]]
In [152]:
print(classification_report(Y_test,rfc_pred))

              precision    recall  f1-score   support

           0       0.87      0.97      0.92     33115
           1       0.83      0.49      0.62      9543

    accuracy                           0.86     42658
   macro avg       0.85      0.73      0.77     42658
weighted avg       0.86      0.86      0.85     42658
In [ ]:

GradientBoosting Classifier¶

In [153]:
from sklearn.ensemble import GradientBoostingClassifier
gbc = GradientBoostingClassifier()
gbc.fit(X,Y)
gbc.score(X,Y)
Out[153]:
0.8495636212753089
In [154]:
cv3=cross_val_score(gbc,X,Y)
print(cv3)
print()
print(cv3.mean())

[0.84967826 0.84851788 0.85206934 0.84225332 0.84288628]

0.8470810164886384

Model Evaluation¶

In [159]:
from sklearn.metrics import confusion_matrix,classification_report
gbc_pred = gbc.predict(X_test)
In [160]:
RocCurveDisplay.from_predictions(Y_test,gbc_pred)
plt.show()
In [161]:
print(confusion_matrix(Y_test,gbc_pred))

[[31434  1681]
 [ 4724  4819]]
In [162]:
print(classification_report(Y_test,gbc_pred))

              precision    recall  f1-score   support

           0       0.87      0.95      0.91     33115
           1       0.74      0.50      0.60      9543

    accuracy                           0.85     42658
   macro avg       0.81      0.73      0.75     42658
weighted avg       0.84      0.85      0.84     42658
In [ ]:

XGBClassifier¶

In [163]:
from xgboost import XGBClassifier
xg = XGBClassifier()
xg.fit(X,Y)
xg.score(X,Y)
Out[163]:
0.8796846539562426

Model Evaluation¶

In [164]:
xg_pred = xg.predict(X_test)
In [165]:
RocCurveDisplay.from_predictions(Y_test,xg_pred)
plt.show()

Results and Conclusion:¶

Best Models in terms of accuracy (In my Experiment):¶

1) XG Boost Model¶

4) Decision Tree Classifier¶

5) Gradient Boost Classifier¶

Best Models in terms of Computation Time (In my Experiment):¶

1) Logistic Regression¶

2) Decisoin Tree Classifier¶

3) XG Boost Model¶

Conclusion:¶

The Ensemble model XG Boost is performing well when compared with Other models like Logistic Regression,Random Forest,Decision Tree and GradientBoost. But XG Boost model consumes more time to train the model.¶

In [ ]: